Sensory Perception & Interaction, Research Group

Human emotions are complex, dynamic and indi- vidualistic experiences that evolve over time, making it chal- lenging for devices to accurately estimate personal emotions in real-time. Traditional emotion models that assume generalized emotions exist as discrete states fail to capture the valuable information inherent in the dynamic nature of emotions. We present an interface that supports multi-resolution emotion self- reporting procedures, demonstrating the construction of emotion labels along a custom emotion scale at the rate of emotion change. This procedure differentiates not only what the emotions are, but also how they are transitioning, enabling the identification of emotions such as “hopeful but getting stressed” vs “hopeful and starting to relax”.